Exploring new possibilities with CDI metrics

- [Kevin] Hello everyone, and thank you for joining today's webinar, leverage AI-powered CDI coding technology to support HCC and risk adjustment. My name's Kevin Cooper, the Director of Product marketing with Optum, and I'll be your host today. Before we begin, please note the following housekeeping items. At the bottom of your audience console are multiple application widgets that you can use to customize your viewing experience. If you have any questions during the webcast, you can click on the Q&A widget at the bottom of your screen to submit a question. We do capture all questions, and will be providing follow up to questions as appropriate. If you experience any technical difficulty, please click on the help widget. It covers common technical issues. You can expand your slide area by clicking on the maximize icon on the top right of the slide window, or by dragging the bottom right corner. There's also a survey widget, which you can use at the end of the webcast to provide us feedback on today's presentation. Additionally, this presentation uses streaming audio. You may listen to the audio through your computer speakers or your headphones. To ensure the best possible system performance, please be sure to shut down any VPN connections and connect directly to the internet. Presenting today will be Tammy Lander, senior manager of CDI and inpatient coding at University of Pittsburgh Medical Center. Tammy's worked with UPMC for over 30 years. In her current role, she provides oversight of the clinical documentation improvement program and inpatient coding for hacks PSI's mortality teams. And Michele Maier, she's the associate director of strategic solutions advisors at Optum provider market solutions. Michele has over 20 years of experience in the healthcare industry and serves as an analytics strategic solution advisor for Optum Enterprise CAC and CDI technology. She actively works with many hospital organizations presenting analytics and consulting on best practice utilization. With that, I'll turn our agenda over to Tammy. Thank you.

- [Tammy] Thank you, Kevin. I'm very happy to be here today to share our journey through the HCC process. What Michele and I will be talking about today are some of the key organizational drivers for change that led to our reviews of HCCs, how we leverage AI-powered technology solutions to transform the revenue cycle, how we augmented our current CDI and coding workflow processes, how we took advantage of the BI analytics to enable performance monitoring and impact, and then we'll share with you our key success takeaways, and at the very end you'll have the opportunity for question and answers. A little bit of background on UPMC, so we are, as Kevin said, the University of Pittsburgh Medical Center, located in Pittsburgh, Pennsylvania. We are a $26 billion healthcare provider and insurer. We have 95,000 employees, over 40 hospitals, 800 doctor's offices and outpatient sites, and our flagship hospital, UPMC Presbyterian Shadyside located in Pittsburgh, ranks annually in several specialties by the U.S. News & World Report, and additionally, our UPMC Children's Hospital of Pittsburgh is consistently in the top 10 children's hospitals as ranked by the U.S. News & World Report. We also collaborate with the University of Pittsburgh School of Medicine, as well as the School of Health and Rehabilitation Sciences.

- [Michele] Welcome everyone. This is Michele. Optum is partnering with our customers like UPMC to transform the middle revenue cycle, focusing on more efficient workflows and accelerating automation on specific tasks to free up your staff to concentrate on more complex needs and analysis, ensuring staff has the tools to proactively work potential edits, supporting that clean claim submission to prevent any claim rework, leveraging data as that strategic asset for resource allocation and evaluation and monitoring of performance, deeper level of support for risk adjustment use cases, so you're successful in those risk-bearing contracts and performance monitoring.

- [Tammy] Some of the key organizational drivers that led us to this change, just to give you a little bit of background, our CDI program started in 2013. We started out with only five of our UPMC hospitals, and we had eight CDI specialists. At that time we partnered with Optum, and we were utilizing their CAC, and we had integrated their CDI program as well. That takes us to today, where we now have 23 hospitals in the corporate revenue cycle, as well as we now have 47 CDI specialists and one CDI manager, and so many of those happened through various integrations as our health system grew. One of the things that happened is that the executives came to us and wanted to integrate the HCC reviews and risk adjustment reviews into the concurrent CDI program, and so historically our focus with our CDI reviews has been on the financial impact of the documentation, ensuring that we're capturing our major comorbidities and comorbidities, but as many of you are well aware, our clinical documentation improvement programs across the country have evolved from that, and so in addition to ensuring that we're capturing any financial opportunities, we want to ensure that we're capturing the severity of the patients that we treat, so how sick they are, and being an academic medical center at our flagship hospital, that's key to show just how very sick these patients are, just beyond the acute care that we're providing, and so our executives came to us, and we're hoping that we can integrate the HCC reviews into our concurrent reviews as well, and so one of the drivers there is to, as I said, ensure revenue integrity, and as UPMC is also an integrated healthcare delivery system model where we have our own health plan, so that was critical into supporting our partner with the health plan as well. The other key takeaway is to ensure our quality rankings, and that by capturing the HCCs and other risk adjustment diagnosis, we can help support our quality team and our quality rankings.

- [Michele] So we're gonna go to our first poll question this afternoon. Do you currently have an integrated Enterprise computer assisted coding and CDI system product And we'll give everyone a minute or two to go ahead and answer this question. The options are yes, you have an integrated computer assisted coding and CDI tool, a computer assisted tool only, a CDI tool only, no to CDI or eCAC, no, but likely to explore in the next one to two years, or the final selection is I don't know. And again, we'll give everyone another minute or so to complete that first survey question.

- [Tammy] Michele, as I mentioned, UPMC, we currently have the integrated Enterprise computer assisted coding product, as well as the CDI product, which is very helpful, and it allows our CDI staff to work very closely with our coding staff and vice versa. Having that integrated process lends itself to a lot more efficiency between the two areas, and really allows for complete coding.

- [Michele] Agreed. In my previous jobs, I've seen independent ones versus the integrated one, and definitely do agree that the integrated one is a better solution in my mind. So we'll give everyone another few seconds to answer the survey. All right, with having everyone had a minute or two to answer, we actually do see that a majority of the respondents do have an integrated enterprise computer assisted coding and CDI tool similar to UPMC, at almost 34%. Right behind that is actually probably a combination of the two, either having a CAC tool or a CDI tool. Again, no is reasonable at about 15%, 7%, almost 8% are in the likely to explore in the next year to two, and then the last is 14%, is the I don't know, so again, that majority is similar to UPMC, which is great. One thing from a strategic partner perspective here at Optum, our enterprise CAC and CDI tool really does utilize the model of the voice of the customer, so as UPMC expanded or looked to expand their CDI program, we were able to work with them to show them how our current functionality could actually fit into their current workflow and show those additional CDI markers or those CDI flags, or the CDI flags for those HCC conditions, to continue to support their goals. And then we continuously will partner with our clients on development of other risk adjustment models, which is one of the next goals that UPMC has, such as the Elixhauser model. Looking specifically at the AI-powered technologies that we look at from an Optum perspective, it kind of goes to our vision, especially for middle revenue cycle, where we look at charging, coding, CDI, but we're also looking to, looking at both inpatient and outpatient, professional and facility, so all of those areas, and then using analytics and auditing and compliance to kinda see how we can improve those functionalities, especially in our artificial intelligence technology solutions, to continue to drive the best from a middle revenue cycle perspective going into the future. This will lead us into our next polling question, which is, does your CDI Enterprise CAC product review for HCC or risk adjustments? And the options include, AI driven, so artificial driven HCC capture, yes, but it's a manual process to capture those HCCs, no, HCC review is currently out of scope, or the fourth answer, which is you are not sure, and we'll give everyone, again, another minute or two to answer this question.

- [Tammy] So Michele, what I like about the Optum tool in regards to this, and as I alluded to, of course, you know you have the HCC markers available in the product, which assists the CDI specialist when they're reviewing to look for those opportunities, but also from a coding workflow perspective, Optum identifies risk adjustment diagnoses in the code summary, and so you're able to auto sequence those codes into the top, and prioritize those codes, and identify which ones are a risk adjustment code, so that's extremely helpful, as well as which ones are specifically HCC adjustment codes.

- [Michele] Agreed. As I've done some research, it is always interesting to see how many of those HCC conditions actually are also kind of, in my mind, almost kind of play a multi or dual purpose, 'cause they have impacts from a risk adjustment perspective in other models as well. So we can go over the answers to this question now, and that is that 25% actually do have artificial driven, artificial intelligence driven HCC capture, which is great, and right behind that, actually, at the largest volume actually, is no, HCC review currently is out of scope for about 30, almost 30%, which is interesting. And then another 20% at yes, that manual HCC process is currently there today. So shifting to kinda talk through, from an Optum technology solution perspective, what does that HCC or risk adjustment workflow look like? There's a number of elements that play in to each of the areas. One thing to remember, though, this is an integrated solution, so what the CDI specialists do and can see during those concurrent reviews, the coder can see post-discharge while they're doing their coding, and then long term into the reporting solution, we're able to utilize that data for reporting purposes. Starting off from a CDI perspective, we leverage clinical language intelligence to review all documentation associated with each case, to look for those gaps, look for those gaps so that UPMC's team, as Tammy mentioned before, could look for opportunities not only on those HCCs, those hierarchical condition, those chronic conditions, but also CCs and MCCs, and other risk adjustment opportunities. And then to quantify those documentation clarifications, our technology actually tracks shifts not only in that traditional sense for revenue integrity, but also quantifying the impact of capturing an additional HCC condition, or clarification to a more complex HCC condition using the HCC factor score, and we calculate this by adding up the weights of each HCC condition category that is present on a case, which long-term will impact that patient's overall risk score. And then kind of moving to that coding summary side from a coder perspective, but again, all users are able to see this, you can actually see that HCC factor score throughout each review, each coding session, even when auditing is doing it post-discharge. You can see that through each step. Also, in the code summary, each HCC is indicated so you know which codes are HCCs. You also know what codes are also risk adjustment for other models. In addition, we automatically will sequence those hierarchal condition categories higher. One thing from a coder perspective, which is what I was many, many years ago, it's nice to see the system automatically sequencing those higher, 'cause we know we want them on the claim, but to think that I'm gonna remember of, what, close to 3000 codes, to make sure those ones are being sequenced higher up, it's nice to see the system automatically doing it for us. In addition, the users can actually review, and we call it HCC lookback, so they can actually look to see, last year and this year, what other HCC conditions have been documented and coded on previous cases for a patient, so to see, has that documentation and the coding been continuous on those patients' records, year over year, as we need them to be? And then in our reporting solution, we have operational dashboards and reports to track the impact, not only of UPMC but their staff performance on those reviews, and then looking to ensure that we can actually have ad hoc reporting to customize the data to meet UPMC's goals and needs to show the success of their initiatives. Then through our performance management program, which is PMP, that allows us to actually collaborate directly with UPMC leadership to track those goals and to monitor new initiatives, ensure that they're seeing success in those, and if there is a, if there is a negative when we're expecting to see a positive, working to kinda help to see what action plans may be appropriate to be put into place. The next poll question we have for today's phone call is, your Enterprise CAC CDI product, does it provide ad hoc and customizable BI analytics reports and dashboards? One, yes, it has ad hoc BI analytics tools, which means you can customize and kinda create your own reports. Yes, canned reports, what others may call standard reports, do you just have standard reports available to you? No, your CAC CDI product has no tool reporting capabilities. And then the last is that you're not sure.

- [Tammy] So as Michele will tell you here in a few minutes, EPMC and Optum, of course, does have the ad hoc BI analytics tool, which has been very helpful in the implementation of this program. It really helps us to identify our opportunities, as well as dig down a little farther. If we have certain physicians we wanna educate, we're able to identify those physicians. It really gives us a lot of flexibility, and Michele's been a great partner with our organization in providing that support as well, and helping us implement these reports, and if we're not sure how to run a report, we can always contact Michele, and she helps us through the process, so that's very, very nice to have that support.

- [Michele] We'll give everyone another minute or so to answer this question before we move on. As the, I wanna call it sometimes the data analyst in me, while I started off my life as a coder, kind of, I really do enjoy being able to see that data and help clients such as UPMC kinda dive into that data to see what's available, so whether you have ad hoc or canned reports, I think any data to kind of help make sure that you're seeing success in a new initiative is always important. So starting to look at the reporting capabilities, a majority do have ad hoc analytics tools, which is great, about 38%. Right behind that is the I'm not sure, which is okay, and then about 19% are saying no BI tools or reports, and then about 8% with canned reports only. So again, a good distribution, but I'm happy to see that most everyone does have, or at least the majority does have those ad hoc BI analytic tools available to them.

- [Tammy] So next I'm gonna talk about how we integrated the HCC reviews into our workflow. Obviously, when we implemented the HCCs, we were not adding staff to our team, and so we had to make sure that we were maximizing the efficiency of our core workflows, and so we were able to integrate the HCC workflow into our existing MS-DRG and APR, the APR-DRG based reviews, and our focus is on inpatient only, so even though we're inpatient cases and HCC is more of an outpatient methodology, it's important to capture these diagnoses at least once a year if they're in the hospital, and that helps support the HCC from the physician's standpoint. The next part of the workflow was to accelerate the automation, so we were able to leverage the evidence-based HCC clinical markers that Optum has to gain stakeholder buy-in, such as the CDI specialists, the coders and physicians, and then finally we had to make sure we had deeper support for the risk adjustment, and so the use of the HCC score within the Optum tool helps us to show the value of the work effort, and to help us to sustain the continual process improvement. So moving on to some of the workflow changes, the first step was to look at the software training, so as I mentioned, we use this tool on the inpatient side currently, and so we have the advantage of the Optum strategic service advisors to support us in system training on the HCC functionality in the CDI tool, and we were able to choose when we implemented the markers, so with the tool that Optum offers, you are able to determine which markers you want to utilize, and if there's any that you wouldn't you have that option as well, so as Optum introduced the HCC markers, we had made the decision to implement them in smaller chunks, so to speak, and so Optum currently has seven HCC markers, and we decided that we would ease our team into them, so we broke them down into two or three, groups of two or three over the course of a few months, and so that allowed us to provide additional education. Additionally, through the training, Optum offers a take a break, as well as quarterly user group sessions that provide continual updates on HCC enhancements, as well as other types of education. In terms of the HCC clinical education, Optum provides monthly clinical education that's available to their clinical documentation improvement specialist coders. Anyone who uses the tool is able to participate in those. Additionally, they also offer a risk adjustment education, and as Michele mentioned, HCC is a specific methodology, but what you will find as you move into different types of methodologies, such as Elixhauser and various risk adjustment methodologies, there is a lot of overlap with the risk methodology adjustment, so integrating the HCCs, we really had nothing to lose by implementing those. We were helping support our health plan, as well as many of those HCC markers are beneficial from other risk adjustment methodologies to show, whether it be our APR severity of illness or, as I mentioned, Elixhauser, which a lot of programs are now using for benchmarking. Next we looked at our focus areas, so there are over 9,000 HCC codes, which break down into 79 categories. If you were to go and introduce that to your team, that would be extremely overwhelming, and so what we decided to do in combination with Optum is to focus on the more common categories, and so what we did is, as I mentioned, we broke down those markers into groups and introduced them to our team over the course of a few months, and so that allowed our team to get used to those HCC query markers, and be able to validate them and integrate them into our workflow, so that really helped optimize our workflow, because we had the education that was provided by Optum. We had the markers which were, again, there, and we were able to enable them with the help of Optum, and so that really helped to build that support, the awareness, and the collaboration, between ourselves and Optum, and it really helped to support our team. Additionally, one of the things we needed to do was to implement policies and procedures to ensure best practice adherence, because sometimes with the risk adjustment diagnoses you have to make sure that you're still following your coding guidelines so that it meets that meet criteria, the monitored evaluated treatment, and falls in line with the coding guidelines. And then finally, as Michele talked about with the polling question, having the reporting in place to monitor the process and identify the impact. The Optum BI analytics custom reports, we continue to develop and enhance them with the help of Michele, as well as a lot of organizations such as ourselves, we do have our own support team from, in terms of being able to run reports, so between the two it really enhances our workflow. And then finally we meet monthly with Optum and Michele through our performance monitoring program, and so Michele is able to customize these monthly reviews, looking at if there's any particular executive insights or trends that we wanna look at. She can help us support those requests and make sure that we're giving feedback to our executives in a timely manner, and really to help, since our focuses shift beyond the financial impact of our reviews and branching out into that quality and risk adjustment, through the tools that we have available in Optum we're able to show that value through the mechanisms, excuse me, the mechanisms they have in place in their tool, and they're subsequently able to report on those.

- [Michele] So kinda speaking to those BI analytics or to the monitoring, the impact, from a change management insight perspective, we as Optum share from a CDI and coding analytics perspective, working to kind of work with our clients to optimize those workflow for that value-add insightful HCC data, empowering our operational leaders with those robust reports to make sure that they can make changes, and then supporting those deep dives through customization to kinda meet those organizational needs and goals, since each organization can be a little different, and ensuring, as Tammy mentioned, if there's a request from UPMC leadership that those operational managers have the answers to kinda follow back up with and then can be able to continue to kind of track and trend, so when we talk about key performance indicators or KPIs, especially when it comes to HCCs, what, when, and how. So the first is what. What metrics do you wanna track? What should you be tracking? What are you targeting? I think usually where I start with with most of my facilities and my organizations, especially like UPMC, okay, what are you looking to do? What's that organizational goal? What did success, what would be defined as success? And then making sure that we not only look at that outcome KPI or that success measurement, but also looking at process ones as we go along, such as how many times did the staff query or identify those documentation gaps for those HCC conditions? And then looking even at those queries to say, what was the physician response and agreement rates from those queries? Since we know, as you expand out your program, you're expanding out and kind of reaching providers in a different way than you did historically, is that having an impact on some of those other metrics that you have, those KPIs you have been tracking year over year? And then never forgetting to understand your baseline data. Without knowing where we started, we can't easily track how far we've come or where we plan on being able to get to, so understanding what we wanna track is great. Then the question becomes when. You wanna definitely do it frequently and continually. I always find, with my process metrics, you probably wanna do it a little bit more often, such as weekly or monthly. I know initially with UPMC when they started off in those focus areas, we kinda focused in on those first few focus areas first to say, okay, here's the gains you're seeing in those queries and here's what that's looking like. Again, UPMC has, I think Tammy mentioned 23 facilities. Where were we seeing those opportunities first found? Were we finding more queries at some of the smaller hospitals versus the larger hospitals, and being able to kind of look at that information and understand it, and then adapt what processes, what workflows were in place, based off of those data trends and those insights that we were finding when looking at those process metrics on that more frequent basis. And then also not forgetting to look at your outcome metrics, usually monthly or quarterly, so that you have that long-term plan, not only to say, yes we had success, but to ensure that you have sustained improvement and ability to kind of change with times as you start to see shifts in numbers. How? How do you wanna track them? So again, definitely BI analytics tools are excellent, but kinda using the tools you have available to you, and then not only using those tools but making sure that you develop those action plans or resolve any outliers that you see. Things that we've seen over the year or that we, I'm anticipating that we will see, we know just recently we had the 2024 model for HCC being finalized. Okay, what shifts do we anticipate as that model comes in? So starting to look at what kind of operational action plans might we need to put into place? So let's start by looking at what I will call a process KPI. We specifically initially looked at kind of two metrics here, average number of HCC codes per case, so again, not that how impactful is a particular HCC, but just knowing, was an HCC condition present at all on a case, and kinda tracking that over the past few years, especially as UPMC started their journey, and then also looking at the number of queries that were sent specifically for those HCCs, so you can kind of see, over the past few years, UPMC has actually seen an increase from their baseline of 2020 to 2022, a 54% increase in that average volume of HCC codes per case, again, not only going back to those queries, but also ensuring that everyone, both in that integrated solution, both the coders and the CDI, saw the value of those codes and saw the impact that they were having. One of the other things that I find a lot of times, especially with organizations as they start on a new initiative, is making sure that we continuously continue to track metrics that we had been tracking, so for example, risk of mortality. As Tammy mentioned, part of why they did this wasn't just for that HCC from a Medicare HMO, Medicare Advantage perspective, but also looking from a quality perspective, so also wanting to track risk of mortality, so we've tracked year over year changes in risk mortality, and seeing here you can see that they've seen a 3% increase from baseline of 2020 to 2022. This insight here from a revenue integrity perspective is really looking at that average HCC factor score, so that culmination of HCCs on an individual case, and kind of looking at that average and looking at it monthly, so while we may report out quarterly, a lot of the times we will track on a monthly basis, and that allows us to have those good conversations during that PMP program with leadership, to say, okay, from this quarter to this quarter we're seeing these kind of gains, especially with, with that average HCC factor score, are we seeing increases with it as we saw increases in the number of cases with an HCC factor score shift, so an increase based off of a query or a documentation clarification? And again, UPMC did see an increase in the average HCC factor score from a quarterly perspective, starting with from quarter two to quarter four of 2022. So what's next? As you've seen, UPMC has seen positive results in their process and outcome KPI metrics with the expansion of their CDI program into HCCs. UPMC's continues to track, and we continue to monitor that HCC impact through BI analytics reports through the performance management program. And again, UPMC does have data analysts on their side, so I will commonly collaborate with them to ensure they're seeing the information that they need and that we are seeing the insights, and they continue to move forward with their workflow optimization, both for this current initiative of HCCs, but also into the future as we look at other risk adjustment models, and we continue to partner with UPMC for that long-term sustained improvements, but also improving not only their programs, but continuing to improve our technology as well.

- [Tammy] So some of the key takeaways from our success is that executive sponsorship and organizational drivers are critical for HCC and risk adjustment. You really need to perform a gap analysis on the current state of your CDI and coding programs to identify the key drivers and provide a business case. Really, automation and the integration of the HCCs into the existing CDI and coding workflow is important for staff adoption. Healthcare, you keep asking more and more of your staff members, and so to have that supportive technology and that efficient technology really helps to align your workflow and help with the automation of your workflow. Do more with less. Any time we can achieve those types of successes through the use of technology is a win for everyone, and the one thing with the risk adjustment, it continues to evolve, and we're fortunate to have a great partnership with Optum, and as Michele mentioned, the HCCs, definitely a great start on our risk adjustment, but we continue to evolve and look at other opportunities above and beyond just the HCCs. And then finally, having that realtime and insightful data is essential for your performance improvement. You need to be able to leverage those analytics to monitor the performance and correct the course if needed, and really it continues to help support this initiative from an executive level, as well as it helps to reinforce with physicians, look at your risk scores, look at your HCC scores, and so that really helped us to provide that evidence. As we all know, evidence based information is really critical to getting buy-in, and so having that information has really helped us to get support from all arenas, executives, physicians, staff, et cetera, and so those were really the three key successes and takeaways from our initiative so far with the HCCs.

- [Kevin] Okay, terrific. Thank you, Tammy and Michele. Before we get started with the Q&A portion of today's webcast, just wanna take a moment to remind you of the survey located beneath the slide deck area within your audience console. We'd love it if you take a moment to provide us with some feedback on today's presentation. As a reminder, if you have a question, you can submit it on your question via the Q&A widget located on the left side of your audience console, and with that we'll go ahead and get started with the Q&A.